Vapor Pressures of Binary Liquid Mixtures* - Journal of the American

Soc. , 1941, 63 (6), pp 1752–1756. DOI: 10.1021/ja01851a070. Publication Date: June 1941. ACS Legacy Archive. Cite this:J. Am. Chem. Soc. 63, 6, 175...
1 downloads 0 Views 458KB Size
J. R. LACHER AND ROY E. HUNT

1752

compared with quantum mechanical calculations. Melamine monohydrochloride hemihydrate has been briefly examined and found to be orthorhom-

[COXTRIBUTION FROM THE

VOl. 63

bic, a = 16.75 A., b = 12.29A., c = 6.93 A., space group, Cmcm,Cmc or CZcm. PASADENA, CALIFORNIA RECEIVED MARCR28. 1941

METCALF CHEMICAL LABORATORY, BROWNUNIVERSITY]

Vapor Pressures of Binary Liquid Mixtures* BY J. R.LACHERAND ROY E. HUNT A measurement of the total and partial pressures of a binary liquid mixture leads to information concerning the chemical potentials of the liquid phase components provided certain thermodynamic properties of the gaseous phase are known. A knowledge of these chemical potentials is important for the development of theories concerning the structure of liquids and liquid mixtures. I n the present paper are reported measurements on mixtures of chlorobenzene-ethylene bromide and 1-nitropropane-ethylene bromide.

Fig. 1. Experimental Details.-The apparatus, which is quite similar t o that used by Sameshima,’ is shown diagrammatically in Fig. 1. Instead of boiling the liquid in B by means of an electrically heated wire, a small pump A, of dimensions recommended by Swietoslawski’ was used t o force liquid and vapor over a thermometer well in B. This well was made quite long in order t o permit the detection of any temperature gradients in B. A and B were

* Original manuscript received December 6. 1939. (1) J. Sameshima, THIS JOURNAL, 40, 734 (1918). (2) “Ebulliometry,” W. Swietoslawski, Jagellonian University Press, 1936.

heated electrically. The vapors in B passed up a heated tube of large diameter, condensed, and returned t o the reservoir via a hold-up trap C. The drop counter below C indicated the speed of circulation of the vapor. The heated parts of the apparatus were insulated with asbestos and shielded with aluminum foil. C and B are fitted with mercury sealed ground glass caps and liquid samples were removed for analysis in ice-jacketed pipets. A cooling coil connected the still to a manostat and pump. Most of the condensation occurred in the water cooled condenser, but the cooling coil was necessary to prevent a slow diffusion of vapor out of the still. The manostat had a 45 1. capacity and was thermally insulated. Dry air was used as the confining gas. Pressures were read with a mercury manometer using a calibrated paper scale. The individual pressure readings could be made t o 0.2-0.3 mm. However, all the pressures reported in this paper represent the average of a large number of individual readings. The pressure in the apparatus did not drift provided the cooling coil was maintained at a sufficiently low temperature and that the temperature of the manostat did not change. A tenth degree mercury thermometer read with a good telescope permitted temperature measurements accurate to a few hundredths of a degree. Experiments performed on pure liquids and mixtures of ethylene bromide-chlorobenzene and ethylene bromide-lnitropropane (whose components have almost identical vapor pressures) showed that the boiling temperatures were independent of a three-fold variation in electrical energy supplied to pump A. There was a long region of constant temperature in the upper part of the thermometer well. The temperature a t the bottom depended on the heating current supplied to B. For most runs this current was adjusted so that all temperature gradients along the well vanished. However, their presence has no effect on the final experimental data which, for the above mixtures, satisfy the Margules equation. When mixtures of ethylene chloride-sym-tetrachoroethane were boiled in the still, the temperature of boiling increased with increasing pump speed and reliable measurements could not be made. The vapor pressures of the pure components differ by almost a factor of ten in this case. Ethylene bromide-sym-tetrachloroethaneand chlorobenzene-sym-tetrachloroethane mixtures are made of components whose vapor pressures differ by a factor of two; their boiling temperatures were independent of the pumping speed. However, a comparison of the data with the requirements of the Margules equation showed that the mole fraction of the more volatile component in the vapor

VAPORPRESS~USS OF BINARY LIQUID MIxrums

June, 1941

phase was too small, giving rise t o errors of 1-2 in the per cent. deviations from Raoult’s law as calculated by the method of Gilmann and Gross.’ The present apparatus will give reliable results for liquid mixtures whose components have almost identical vapor pressures. However, it is not effective in producing equilibrium between the liquid and gaseous phase when they differ considerably in composition. In such cases it is better to pass the vapor through a liquid mixture of the same components as suggested by Rosanoff, Lamb and Breithut* and as is done by Scatchard and his co-worker~.~ Technical chlorobenzene, ethylene bromide, and 1nitropropane were dried over calcium chloride and fractionated repeatedly in a n efficient still. The densities of the liquid mixtures were determined in Weld specific gravity bottles having a capacity of 5 cc. The results, obtained a t 30°, are expressed by the equations

+ 1.0635N~- (0.174 + 0.042NB)NBNC dz = 0.9904 + 1.1693N~- (0.0495 4-0 . 0 0 8 0 l V ~ ) N ~ N ~

1753

TABLE I1 ETRYLENE BROMIDE-~-NITROPROPANE P mm.

YB

NB

0

1 0.9137 .8367 .7254 .6352 .5219 .4380 .3259 .2133 .0873 .0539 .0468 0

119.7 123.9 127.4 130.9 133.0 133.0 132.0 128.9 124.4 119.8 117.6 116.7 115.0

0 20.9 41.0 65.0 80.8 85.0 80.9 65.4 42.6 21.5 11.7 6.4 0

1 0.8842 .7622 .6627 ,4726 .2754 .1650 ,0906 ,0295 0

(b) 120’ isotherm 549.5 0.8435 581.0 .7220 599.8 ,6365 609.5 ,4829 612.7 ,3153 600.3 ,2054 586.1 ,1228 571.5 ,0419 557.4 0 550.2

0 37.3 63.7 77.9 82.7 64.3 44.6 25.3 8.2 0

1 0.9431

.so1

.ma

.6562 .5085 ,3995 ,2698 .1580 .0575 .0352 .0298

di = 1.0959

(1) dl and dz represent the density of mixtures of ethylene bromide with chlorobenzene and with 1-nitropropane, respectively. NB, Nc, N N give the mole fractions of ethylene bromide, chlorobenzene, and 1-nitropropane. Because of the large difference in density of the pure components and the low volatility of the mixtures a t room temperature, the composition could be determined quite accurately.

Experimental Results.-Table I gives the data obtained by measuring pressure-composition isotherms a t 75 and 100’ for ethylene bromidechlorobenzene mixtures. Such mixtures will be called system one. Table I1 gives similar results for ethylene bromide-1-nitropropane, system two, a t 75 and 120’. Measurements a t 100’ TABLE I ETHYLENEBROMIDE-CHL.OROBENZENE NB

YB

P,mm.

FE,cal.

1 0.9214 .7895 ,6443 .5132 .3576 .1684 0

(a) 75’ isotherm 1 119.9 0.9034 122.3 .7633 125.6 ,6268 127.7 ,5101 128.4 .3707 128.0 .1882 125.4 0 121.9

0 11.4 28.2 38.6 41.4 37.2 20.3 0

1 0.9207 .7901 .6434 .5129 .3568 .1693 0

(b) 100’ isotherm 1 295.3 0.9053 300.3 .7695 306.4 ,6306 310.0 .5130 311.2 .3707 309.5 .1%1 304.4 0 296.1

0 11.0 25. G 34.4 37.2 32.5 19.5 0

(3) H. H. Gilmann and P. Gross, THIS JOURNAL, 60, 1525 (1938). (4) Rosanoff. Lamb and Breithut, ibid., Si, 448 (1909). (5) Scatchard, Raymond and Gilmann, ibid., 60, 1275 (1958).

FE,cal.

(a) 75’ isotherm

1

for system two are not reported since they furnish no further information; except for this, all runs are included. N represents the mole fraction of the appropriate component in the liquid phase, Y the mole fraction in the gas phase, P is the total pressure in mm. mercury, and FE represents the excess free energy of mixing in cal./mole. In order t o compute activity coefficients from the experimentally measured quantities, i t is necessary t o take into account the deviations from the gas laws. The equations used are given by Scatchard and c o - w o r k e r ~and ~ ~ ~are for the system ethylene bromide-chlorobenzene

PYC

log Yc = log P\ Nc

+

(Pc

-

VC)(P 2.303RT

- PL)

Po and V refer to the vapor pressure and molar volume of the liquid component. The P’S are constants depending only on temperature and express the fact that the vapors do not obey the ideal gas law. ,Bc was calculated from the theory of corresponding states using the critical temperature and pressure listed in the “International Critical table^."^ PN was taken equal (6)Scatchard, Wood and Mochel, J . Phys. Chem., 43, 119 (1939). (7) “International Critical Tables,” Vol. 111, p. 245.

to the measured value for nitromethane.8 PB was estimated t o be one-half as large as pc by comparing van der Waals constants for ethylene bromide and chlorobenzene. Since P - Po for these systems is quite small, the uncertainties in the p’s have very little effect on the values of log 7. The maximum correction in log y was 0.0005 for the ethylene bromide-chlorobenzene system and 0.0011 for the other mixture. If the experimental data satisfy the requirements of the Margules equation, the activity coefficients should be expressible by equations of the forms

In order t o illustrate the accuracy with which these a’s will reproduce the experimental data, we have listed in Table IV the observed and calculated partial pressures, P Y , a t 75’. The data TABLEI V EXPERIMENTAL AND CALCULATED P A R T I A L 75O I N MM.

.v B 0.9214 ,7895 ,6443 .5132 ,3576 .1684

PYB

Calcd.

P YB Exptl.

PKESSLrRES A T

PYC Calcd.

P1-r Exptl.

(a) Ethylene Bromide-Chlorohermne 110.7 110.d 11.9 11.8 95.9 95.9 29.9 29.7 80.0 80.0 17.8 41.1 65.5 63.0 62.9 65.5 47.6 47.4 80.5 80.6 23.8 23.6 102.0 101.8 (b) Ethylene Bromide-1-Nitropropane

The values of CYZ and CYS which will reproduce the activity coefficients are listed in Table 111. TABLE I11 (a) ETHYLENE BROMIDE-CHLOROBENZENE Temp.,

O C .

75 100

a2

a8

0.1598 .1115

0.0993 .1250

(b) Ethylene Bromide-1-Nitropropane 0,2795 0.3000 75 .2315 ,2834 120

140

0.9431 ,8801 .7648 ,6562 ,5085 ,3995 ,2698 .1580 .0575 ,0352 .0298

113.1 106.4 94.8 84.3 69.6 57.9 42.2 26.5 10.3 6.4 5.4

113.2 106.6 95.0 84.5 69.4 57.8 42.0 26.5 10.5 6.3 5.5

10.6 20.7 36.3 48.6 63.5 74.0 86.5 97.8 108.6 111.1 111.6

10.7 20.8 36.0 48.5 63.6 74.2 86.9 97.9 109.3 111.3 111.2

together with the total pressures, are graphed as a function of N B in Figs. 2 and 3. The dots represent experimental and the continuous curves calculated pressures. The agreement is excellent.

100

4

E

a;

60

20

0.4 0.8 Mole fraction, Nn. Fig, 2.-The 75’ isotherm for ethylene bromide-chlorobenzene. Points represent measured total and partial pressures; continuous curves satisfy the Margules equation. 0

(8) Eucken and Meyer, Z. physik Chem., B6, 452 (1929). (9) Hildebrand, “Solubility,” Reinhold Publishing Corp., New York, N.Y . , 1936.

Mole fraction, N B . Fig. 3.-The 75’ isotherm for ethylene bromide-lnitropropane. Points represent measured total and partial pressures; continuous curves satisfy the Margules equation,

June, 1941

VAPORPKESSURES OF BINARY LIQUIDMIXTURES

1753

The excess free energy of mixing was calculated from the equation FE = 2.303RT[N~log YB f N e log r e ] =

[: + T ( 2 - No)]

2.303RTN~Nc

(4)

The values of FE calculated from the directly measured 7's are given in Tables I and I1 and graphed as points in Fig. 4. The continuous curves represent the values of FE calculated from the CY'S given in Table 111. The maximum correction to FE due to gas imperfection was 0.7 and 1.5 cal. for systems one and two, respectively. For both mixtures F E is slightly larger a t 75' than i t is a t the higher temperatures. The maximum difference of 4.6 cal. for system one is, we believe, experimentally significant; the difference of 1.6 cal. shown by the other system is not larger than the experimental errors. In Table V we have summarized some thermodynamic properties for these mixtures a t the mean temperature of our experiments. TABLE V PROPERTIES OF EQUIMOLAL MIXTURES ( v M /VO)100

FE, cal. H , cal. TSE,cal.

Ethylene bromidechlorobenzene

Ethylene bromide1-nitropropane

0.31 39.4 105.6 66.2

0.10

85 98 13

( V'/ V0)lOO is the per cent. volume change on mixing, VObeing the volume of the unmixed components and VM the difference between the volume of the mixture and Vo. TSE was calculated from ( d F E / d T ) , = -SE and H from FE = H -

0

0.4 0.8 Mole fraction, NB. Fig. 4.-Excess free energy of mixing versus mole fraction of ethylene bromide. Upper curves are for mixtures with 1-nitropropane; lower curves for mixtures with chlorobenzene. 0, represent 75' data: 8 , represent 100' and 120' data.

The system chloroform-ethanol is particularly interesting in that i t shows a reversal of sign in the volume and excess entropy change; the excess free energy is positive for all compositions of the mixture. It would be necessary, in order TABLE VI" System

CaHrCsHiz CsHurCC14 CaH~cCli CHCla-GHaOH

(VM/Vo)100

0.65 .16 ,003 .03

TSE,cal.

Mole fraction

101.4 0.5 17.5 . 5

10.7

.5

80 .12 (GHsOH) .oo 0 .28 (GHIOH) - .39 -266 .SO (CzHbOH) We wish to thank Mr. W. B. Buck for making the calculations necessary for the preparation of this table.

to make a quantitative comparison, to have direct measurements of the heat of mixing for the TSE. systems listed in Table V and also, perhaps, for Discussion.-Both these systems show posi- benzene-carbon tetrachloride mixtures. If one tive volume changes on mixing, system one ex- uses the measured value for FE for the latter mixpanding to a greater extent than system two. ture and I'0ld'si3 value of H , then TSE becomes The excess free energy change, however, is twice about 2 instead of 10.7 cal. as great for system two. Our experiments show, The experimental data now available show that we believe, that TSE is greater for system one. no simple relation exists between FE and TS" There appears t o be a qualitative correlation be- but that TS" and ( Vw/17,)100 always possess tween the volume and excess entropy change on the same sign. A possible explanation for the mixing. The accurate measurements of Scatch- latter correlation can be given in terms of the ard and his co-workers on chloroform-ethanol10 probability function W , of Menke.14,1b If two and the three possible binary mixtures from ben- liquids contract on mixing, W must change so as zene, cyclohexane, and carbon tetrachloride6~11~1Y to reproduce the more closely packed arrangelead to the same conclusion. I n Table VI we have ment of the molecules and the excess entropy will listed the relevant data which they have obtained. be negative. The reverse will be true if the (10) G. Scatchard and C. L. Raymond, THISJOURNAL, 60, 1278 (1938). (11) Scatchard, Wood and Mochel, ibid., 61, 3206 (1939). (12) Scatchard. Wood and Mochel, ibid., 69, 712 (1940).

(13) R. D. Vold, ibid., 69, 1515 (1937). (14) H. Menke, Physik. Z.,SS, 593 (1932).

(15) J. H. Hildebrand and S. E Wood, J . Phys. Chcm., 1, 817 (1933).

1756

NOTES

liquids expand on mixing. It does not necessarily follow that TSE will be zero if the liquids are mixed at constant volume, though this might be the case if W had spherical symmetry. We wish to thank Dr. H. C. Eckstrom of this Laboratory for the opportunity of discussing this paper with him. summary ,4n apparatus is described for use in making

Vol. 63

vapor-liquid equilibrium measurements and the conditions under which i t gives accurate results have been ascertained. The excess thermodynamic functions for mixtures of ethylene bromide with (1) chlorobenzene and ( 2 ) I-nitropropane have been measured. Experimental data now available indicate that the excess entropy change and the volume change on mixing possess the same sign. A possible explanation for this is given. PROVIDENCE, RHODEISLAND RECEIVED APRIL12, 1941

NOTES The Mechanism of the Addition of Hydrogen Cyanide to Benzoquinone BY C. F. H. ALLENAND C. V. WILSON

The addition of two molecules of unsymmetrical reagents to benzoquinone almost always gives rise to a substituted quinone (I) in which the new groups are on opposite sides of the benzene ring and in the 2,5-positions. 0

0

+ 2HX I/

0

I OH

OH

OH I1

I11

X

Sometimes the corresponding hydroquinone (11) results. The commonly accepted mechanism' accounts for the formation of this type of product by assuming consecutively 1,4-addition, enolization, and oxidation, the processes being repeated one or more times, depending upon the nature of the reactants. An exception is encountered when hydrogen cyanide is the reagent of the type HX, used as an (1) Gilman, "Organic Chemistry," John Wiley and Sons, Inc.. New York, N.Y.,1938,p. 600.

addend. I n this instance, two molecules add; the product is a dicyanohydroquinone (111),2 in which both CN groups are adjacent and on the same side of the molecule, i. e., unsymmetrical addition has occurred. No intermediates can be isolated; the only other product of the reaction is hydroquinone. The mechanism cannot be the same, therefore, when this reagent is employed. However, an explanation becomes evident when the available unsaturated systems are examined. Addition of the first molecule of hydrogen cyanide takes place as with all HX-type reagents,' and the cyanohydroquinone is oxidized by the excess quinone to form a monocyanoquinone (IV). This cyanoquinone now contains an alternate conjugated system involving the nitrile group, t o which hydrogen cyanide would be expected to add more readily than to thesystem C=C-C=C=O.a The resulting substance would be 2,3-dicyanohydroquinone (111), which is actually obtained. This suggested mechanism is, thus, in accord with the 0 I1

CH I!

0

OH

11

CH /I

CH /c\

CHCN I

C I I

CH

C=C=NH

\./ I/

OH

0

111

known facts about the behavior of conjugated systems. It successfully accounts for the nature of the product and eliminates the apparent ex(2) Thiele and Meisenheimer, Bcr., 38, 675 (1900). (3) Duffand Ingold, J . Chcm. SOC.,87 (1934).